rohitgoel15 wrote:
In the summer of 1936 a polling service telephoned 10,000 United States voters and asked how they planned to vote in the coming presidential election. The survey sample included a variety of respondents—rural and urban, male and female, from every state. The poll predicted that Alfred Landon would soundly defeat Franklin Roosevelt. Nevertheless, Roosevelt won in a landslide.
Which one of the following, if true, best explains why the poll’s prediction was inaccurate?
(A) The interviewers did not reveal their own political affiliation to the respondents.
(B) Only people who would be qualified to vote by election time were interviewed, so the survey sample was not representative of the overall United States population.
(C) The survey sample was representative only of people who could afford telephones at a time when phone ownership was less common than it is today.
(D) No effort was made to determine the respondents’ political affiliations.
(E) Because the poll asked only for respondents’ candidate preference, it collected no information concerning their reasons for favoring Landon or Roosevelt.
OFFICIAL EXPLANATION
(A) No. This would increase the poll’s credibility. Had the interviewers revealed their political affiliations, it would have decreased the poll’s credibility because their opinions might have influenced the people they were interviewing.
(B) No. This would increase the poll’s credibility. The opinions of those unqualified to vote were irrelevant since they did not vote.
(C) Yes. We are looking for something that might have caused the poll to be unrepresentative of the general population. Now, in 1936, phones were not as common as today. If most of the phones were in the homes of the wealthy who held opinions that differed from those of most of the population, then the poll would not have been based on a good cross section of the electorate.
(D) No. Asking for the respondents’ political affiliations would probably help determine how representative the poll was. However, if a disproportionate number of respondents were supporters of Alfred Landon, this choice does not explain why the selection process was skewed.
(E) No. This might somewhat explain the inaccuracy of the poll: Perhaps the support for Landon was weak or very volatile. Nevertheless, there is a much stronger explanation offered.